Digital Media System
Digital Media System:
Extending MX into Digital
A digital programmatic omnichannel system that extends MX into paid social, programmatic display, CTV, and digital media buying. Pixis-class digital capabilities, customised for our agencies, integrated with OOH.
Monday Scoping Session
Miroma AI Team
~200h wall-clock commitment
The Burning Platform
Crowdsauce demoed a modern portal
to Live Nation
£9M/yr
Live Nation account at risk
Meta-heavy
Primary platform for Live Nation
16 team
Sold Out paid social operation
"When we say we've got something coming, it has to be coming, not six months down the road." - Adrian, Sold Out CEO
Crowdsauce's portal showed API-rendered creative previews, targeting breakdowns, and approval tabs. Their capabilities are limited, but the perception created urgency. Drop 1 is the direct counter.
What's Already Built
Meta automation system:
production software, not a prototype
4,998
Lines in core file
(114 functions)
7,300
Total lines
production Python
£122.5K
Annual FTE
savings delivered
~12
Build sessions using
vibe coding methodology
PRODUCTION
Deployed on GCP · Running against Live Nation ad accounts
This system proves the methodology works at production scale. The digital system is a new build: different APIs, different data models, different platform requirements. Estimated independently at ~200h wall-clock.
The Vision
MX OOH + Digital under one AI layer
Nobody else unifies digital and OOH. This is our whitespace.
Agency Workflows
Sold Out
Paid Social · Meta, TikTok, Google
Attentive
Programmatic · DV360, MNTN, Quantcast
SpotCo
Broadway · 15+ vendors, $2.4M budgets
Dewynters
Omnichannel routing · OOH + Digital
↓
One AI Layer
AI Engine
Campaign orchestration · Intelligent pre-fill · Audience & geo targeting · Budget optimisation · Creative validation · Anomaly detection
↓
Platform Features
Approval Portal
Client-facing, one-click
Campaign Builder
Brief → plan → launch
Reporting Engine
Cross-platform aggregation
Finance Integration
Billing, reconciliation, margin
↓
Platform Connectors
Digital Connectors
Meta (DONE) · TikTok · Google · DV360 · MNTN · Quantcast · Amazon · Yahoo · TTD
MX OOH Engine
Pearl · Media owners · Pricing · Bookings
↓
Infrastructure
Shared Infrastructure
Auth (Entra ID SSO) · Database · GCP (Cloud Run, GCS, Firestore) · Sentry · Intruder
Agency Workflows
Three agencies + omnichannel routing
Sold Out
Paid Social
£9M Live Nation account, Meta-heavy
16 team members, 69K+ historical records
~600+ campaigns/month
Meta
TikTok
Google
"They are a client and demand client service." - Paul
Attentive
Programmatic
4,910 line items since 2014
191 unique clients tracked
DV360, MNTN, Quantcast, Amazon DSP
DV360
MNTN
Quantcast
Amazon
SpotCo
Broadway Digital
$2.4M budgets per show (Chess)
15+ vendors, 60+ onboarding tasks per show
Multi-phase lifecycle budgets
Meta
Google
TikTok
MNTN
Dewynters
Omnichannel Routing
Briefs decompose into OOH + digital rows
Routes to MX (OOH) or digital engine
One plan → channel-specific workflows
OOH
Programmatic
Paid Social
Drop 1: The Crowdsauce Counter
Approval Portal + Client Reporting
This is not six months. This is the immediate counter.
→
📊
Dashboard with traffic-light status
→
👁
Creative preview (API-rendered)
→
✅
One-click approve / reject
→
~25h
Wall-clock commitment
Replaces: PowerPoint approval decks · dual manual Box upload · email-based approvals · manual report assembly · fragmented Google Sheets
Platform Connectors
9 platforms in 3 tiers.
Meta is done. Each gets easier.
| Platform |
Tier |
Complexity vs Meta |
Drop |
Status |
| Meta / Facebook |
A |
Baseline |
– |
DONE |
| TikTok |
A |
0.7× |
Drop 3 |
VALIDATED |
| Google Ads |
A |
0.95× ↑0.8 |
Drop 3 |
VALIDATED |
| DV360 |
B |
1.6× ↑1.3 |
Drop 4 |
COMPLEX |
| CM360 |
B |
0.4× NEW |
Drop 4 |
SEPARATED |
| MNTN |
C |
0.2× ↓0.4 |
Drop 4 |
REPORT ONLY |
| Quantcast |
C |
0.25× ↓0.5 |
Drop 4 |
REPORT ONLY |
| Amazon DSP |
B |
0.55× ↑0.5 |
Drop 4 |
VALIDATED |
| Yahoo DSP |
A |
0.35× ↑0.3 |
Drop 4 |
VALIDATED |
| The Trade Desk |
B |
0.65× ↑0.5 |
Drop 4 |
ACCESS GATED |
Tier A: Full automation (campaign CRUD via API). Tier B: Automation with complexity premium (access gated, docs limited). Tier C: Reporting only (no campaign management API exists).
Multipliers validated by API documentation research (March 2026). Each connector is independently estimated based on API complexity.
Delivery Roadmap
Sequential drops,
standalone value at every stage
| Drop | Scope | AI Build | Expected | Commitment | Confidence |
|---|
| Drop 1 |
Approval Portal + Client Reporting |
11.25h |
~20h |
~25h |
Med-High |
| Drop 2 |
Meta Integration + Campaign Build UI |
9.0h |
~16h |
~20h |
High |
| Drop 3 |
TikTok + Google Ads Connectors |
12.5h |
~22h |
~35h |
Medium |
| Drop 4 |
DV360/CM360 + Tier B/C Connectors |
17.0h |
~30h |
~50h |
Low-Med |
| Drop 5 |
IO, Finance, Reporting, Data Isolation |
11.0h |
~19h |
~28h |
Medium |
| Subtotal |
Drops 1–5 |
60.75h |
~107h |
~158h |
– |
| QA Buffer |
0.25× per Ben's directive |
~15h |
~27h |
~40h |
– |
| Total |
Drops 1–5 + QA |
~76h |
~150h |
~200h |
– |
| Drop 6 |
Gen AI Ad Creation (future) |
TBD |
TBD |
TBD |
Unscoped |
Expected = lands here ~50% of the time. Commitment = worst-case-adjusted per drop: P80 for High-confidence drops, P90 for Medium, P95 for Low-Med. Present ~200h as the commitment. The gap is wider for risky drops (Drop 4: +67%) and narrower for proven ones (Drop 2: +25%).
49 sessions across 5 drops. Methodology: ASU v2 (PERT-weighted), validated by API documentation research (March 2026) and estimation research across 22 published studies and 16 classical SE texts. Highest risk: Drop 4 DV360/CM360 (4–10h AI range, gated access). Drop 3 Google Ads gRPC/protobuf is a paradigm shift from REST.
Competitive Positioning
Our whitespace: digital + OOH unified
Cross-platform bid/budget optimisation, AI creative generation. $50M+ ARR, 1,000+ customers. Serves GroupM ($60B) and Omnicom.
Zero OOH capability. No negotiation-based buying. No production workflow.
"World's fastest growing agentic-native system." Creative library, bulk editing, audience targeting, one-click IO.
No OOH workflow. No media owner negotiation logic. No proprietary pricing data.
£1K setup + £2K/mo retainer. Niche event marketing. Modern portal demoed to Live Nation. The trigger for this entire strategy.
No media planning. No OOH. No cross-channel.
Smartly (digital-only), Bionic ($199/user/mo), Pencil (creative-only $14-55/mo), Talon PLATO (OOH-only), Broadsign (DOOH-only), AdQuick (OOH-only)
All single-channel or single-capability. None unify digital + OOH.
Nobody unifies digital and OOH under one AI layer.
87% of UK buyers are increasing DOOH investment. The lines between digital and OOH are blurring.
Requirements Analysis
75 requirements mapped across
two agencies
8 clusters: Briefing & Intake, Campaign Planning, Approval & Activation, Monitoring & Delivery, Reporting & Insights, Client Portal, Internal Team, System Admin. System admin (SSO, security, DR, scalability) is fully shared with MX OOH.
Valuation & Productisation
From agency services to
platform economics
4–6×
Agency Services
Revenue multiple
VS
$10M platform ARR × 10× = $100M incremental enterprise value
Every drop ships proprietary data + IP. The path: internal tool → multi-agency platform → external SaaS.
Drop 1 creates client interaction data. Drop 2 creates campaign performance data. Drops 3–4 create cross-platform data nobody else has. Drop 5 creates financial data. Drop 6 creates creative performance data. Each step increases the valuation multiple.
Frequently Asked Questions
FAQ
How does this relate to MX OOH?
Extension, not separate. Shared auth, database, GCP infrastructure, reporting framework. One platform with media-type views.
What's the total effort?
~200h wall-clock commitment for Drops 1–5 + QA. Expected case ~150h, commitment uses worst-case timings for Low/Medium confidence drops. ASU v2 PERT-weighted, validated by API research and 22 published studies. Drop 6 (Gen AI) scoped separately.
What about MX US?
Separate scope. OOH Drop 4. Architecture supports it but it's a distinct conversation.
Who builds it?
AI-accelerated development using proven vibe coding methodology. Same approach that built the Meta system in ~12 sessions and the contracts project in 3 calendar days.
What about security?
Shared with MX: SSO (Entra ID), Sentry.io, Intruder.io, org-level data isolation. Not double-counted in this estimate.
When does Sold Out get something?
Drop 1 is the stopgap. ~25h wall-clock commitment. Begins after MX Phase 1 foundation (shared auth + DB dependency).
What about Gen AI?
Drop 6, future. To be scoped when core platform is complete and the AI creative landscape has settled.
What if it takes longer than 200h?
The 200h commitment already uses worst-case timings for every risky drop. The only scenario that pushes beyond it is a platform blocking API access entirely (e.g., DV360 denying subscription access). In that case we descope that connector and move on, not add hours. Each drop ships standalone value.
How does this compare to Pixis?
We do everything Pixis does PLUS OOH. They have $209M raised but zero OOH capability. Nobody unifies digital and OOH under one AI layer.
What about cost recovery?
R&D tax credits apply to all dev hours: MX OOH, Digital, and the Meta system already built. 20–33% recovery on qualifying expenditure.
Next Steps
Extending MX
into Digital
✓
MX OOH is priority #1. Paul confirmed. Digital extends OOH, not replaces it.
✓
Digital scoping: done. 75 requirements mapped, 5 drops scoped + Gen AI roadmap, API documentation researched, estimation methodology validated (ASU v2, 22 studies), competitive landscape analysed.
3
Rollout: Paul & Craig → Susie → agencies. Strict sequence per Ben's directive.
4
Drop 1 begins after MX Phase 1 foundation. Shared auth + DB dependency. Approval portal counters Crowdsauce.
5
Live prototype available for walkthrough. Reference the demo for a hands-on view of the platform direction.
Full strategy document available
Detailed estimate document available
Prototype demo ready